Analog Joint Source-Channel Coding with Non-Gaussian Sets of Sensor Samples

This work presents an analog joint source-channel coding scheme as an alternative for machine-type communications that require low latency. We have implemented parametric and non-parametric coding for non-Gaussian sets of samples, collected by sensors of temperature and humidity. Since the parametric mappings are optimal only for Gaussian sources, the Box-Cox transformation was used in order to improve performance. Furthermore, we have employed non-parametric coding, which adapts more efficiently to the data sets, further improving performance at the cost of greater complexity.